RSAC Gandalf Challenge: Insights from the World's Largest Red Team
Join Max Mathys (Lakera’s Software Engineer) and Athanasios Theocharis (Lakera’s Gandalf Engineer) for a live webinar where we’ll look into the latest Gandalf challenge mechanics, showcase the most innovative attack methods, and explore the unique lessons learned about red teaming LLMs.
At this year’s RSA Conference, we introduced an entirely new Gandalf challenge that's not your everyday CTF— it's built around the concept of AI red teaming.
We designed a dynamic environment where successful attacks directly inform defense strategies, meaning once an attack succeeds, all future attempts using similar methods are blocked. This has provided us with invaluable lessons for red teaming AI systems in the real world.
Whether you're a seasoned red teamer or just starting out, this webinar will give you the knowledge and skills you need to stay ahead.
- Look inside RSAC Gandalf: We'll break down the design and goals of the Gandalf challenge, and how it's changing the game for AI red teaming. Plus, some interesting stats from the challenge itself.
- Discover the most interesting red teaming strategies: See the cleverest attack methods that came out of the challenge, and how you can use them in your own red teaming work.
- Learn how the challenge was created: Get a behind-the-scenes look at how the Gandalf team built the challenge. We'll cover the attacks they were targeting, the unexpected roadblocks they hit, and how they adjusted the difficulty.
- See how Gandalf moved to multiplayer: We'll talk about the multiplayer format as well as using crowdsourcing and teamwork to level up your red teaming.
- Meet the surprise guest: In an exclusive interview, the challenge winner will share his first-hand insights from the challenge.
Athanasios Theocharis is a Software Engineer Intern at Lakera and the mastermind behind some of the coolest Gandalf adventures. When he's not brainstorming new levels or coding new features, he can be found immersed in sci-fi books or enjoying a game of table tennis.
%201.webp)
untouchable mode.
Lakera Guard protects your LLM applications from cybersecurity risks with a single line of code. Get started in minutes. Become stronger every day.
Several people are typing about AI/ML security. Come join us and 1000+ others in a chat that’s thoroughly SFW.
